Contrasting Patterns of Genetic Differentiation Between Endemic

Molecular Phylogenetics and Evolution 44 (2007) 474–482 www.elsevier.com/locate/ympev Short communication Contrasting patterns of genetic diﬀerentiation between endemic and widespread species of fruit bats (Chiroptera: Pteropodidae) in Sulawesi, Indonesia

Polly Campbell a,b,*, Andrea S. Putnam c, Caitlin Bonney a,d, Rasit Bilgin e, Juan Carlos Morales f, Thomas H. Kunz a, Luis A. Ruedas g

a Department of Biology, Boston University, Boston, MA 02215, USA b Department of Zoology, University of Florida, P.O. Box 118525, Gainesville, FL 32611, USA c Section of Ecology, Behavior, and Evolution, Division of Biological Sciences, University of California, San Diego, La Jolla, CA 92093, USA d Harvard School of Public Health, Boston, MA 02115, USA e Bog˘azic¸i University, Institute of Environmental Sciences, Bebek 34342, Istanbul, Turkey f Division of Environmental Biology, National Science Foundation, Arlington, VA 22230, USA g Museum of Vertebrate Biology and Department of Biology, Portland State University, Portland, OR 97207, USA

Received 14 July 2006; revised 25 January 2007; accepted 4 February 2007 Available online 22 February 2007

1. Introduction Bridle et al., 2001) implicates evolutionarily recent adaptive radiation. The Indonesian island of Sulawesi is remarkable for its Although, the diversity of Old World fruit bats (Pteropo- complex geological history and high biodiversity. Lying didae) is higher on Sulawesi than on the islands of the Sunda directly East of Wallace’s original zoogeographic boundary shelf and the Malay peninsula combined (Corbet and Hill, between the Oriental and Australian Regions (Wallace, 1992), phylogenetic relationships among pteropodid bats 1876), Sulawesi (Fig. 1) constitutes the largest terrestrial in Wallacea are not well-defined at the species level, and nei- habitat in Wallacea, a biogeographical region prioritized ther broad-scale phylogeographic nor population-level as a global conservation hotspot (Myers et al., 2000). Oce- genetic structure have been examined in any species of bat anic isolation has played a significant role in the evolution from Sulawesi. Here, we evaluate the phylogenetic status of Wallacea’s highly endemic fauna (Whitten et al., 1997); of Cynopterus brachyotis on Sulawesi, and compare the Sulawesi’s considerable size (186,145 km2), complex topog- within-island genetic structure of this widespread fruit bat raphy and diversity of microclimates and habitats have to that of a forest-restricted endemic, Thoopterus nigrescens. provided ample opportunity for diversification and in situ Thoopterus nigrescens (67–99 g) is the sole member of a speciation (Bridle et al., 2001). Molecular studies of Sulaw- genus that is endemic to Sulawesi and parts of the Moluc- esi’s biota have found evidence for both historic vicariance can Island chain (Fig. 1), and is strongly associated with and ecological selection as key factors driving diversifica- primary forest (Bergmans and Rozendaal, 1988). Cynopte- tion (Evans et al., 2004; Bridle et al., 2004). Geographically rus brachyotis (36–41 g) is common in disturbed habitats on concordant patterns of genetic differentiation across dispa- Sulawesi (Bergmans and Rozendaal, 1988) and is the only rate taxa (amphibians and monkeys, Evans et al., 2003a,b) Wallacean representative of a genus whose diversity in pen- support vicariance hypotheses, while low genetic diver- insular Malaysia and the Greater Sunda Islands suggests a gence among ecologically differentiated congeners (shrews, Sunda shelf center of origin (Campbell et al., 2006). Ruedi, 1995, 1998; grasshoppers, Walton et al., 1997; Although the range of C. brachyotis extends from South- west India to Sulawesi (Corbet and Hill, 1992), recent phylogenetic analysis demonstrates that it comprises a complex of evolutionarily distinct lineages, including allopatric lin- * Corresponding author. Address: Department of Zoology, University of Florida, P.O. Box 118525, Gainesville, FL 32611, USA. Fax: +1 352 eages from Sulawesi and the Philippines and two sympatric 392 3704. lineages from the Sunda shelf (C. brachyotis Forest and E-mail address: [email protected]fl.edu (P. Campbell). Sunda; Campbell et al., 2004).

Fig. 1. Map of Sulawesi. Although currently a single island, Sulawesi was formed by accretion of a number of geologically distinct fragments. Major suture zones between fragments are shown in white. Scale bar on main map 100 km. Numbers refer to collecting localities in Table1: 1, Siuna; 2, Poso; 3, Lore Lindu; 4, Mangole Isl. (indicated on inset); 5, Tana Toraja; 6, Buton Isl.; 7, Kabaena Isl.; 8, G. Lompobatang. Inset map localizes Sulawesi (dark gray) and the Moluccas (black) in Southeast Asia. The Sunda shelf (light grey outline) deﬁnes the approximate area of additional terrestrial habitat exposed when sea levels dropped to 6120 m below present levels during Pleistocene glacial cycles. PM, peninsular Malaysia; S, Sumatra; B, Borneo; P, Philippines. Nias Island indicated by star. Scale bar on inset 1000 km.

Within Sulawesi, the taxonomic status of C. brachyotis In the present study, we use mitochondrial DNA remains uncertain; variation in body size has been inter- (mtDNA) sequence data and nuclear microsatellites to preted as evidence that two allopatric populations (Berg- address three main questions. (1) Was Sulawesi colonized mans and Rozendaal, 1988) or two sympatric species by C. brachyotis from different geographic regions, as sug- (Kitchener and Maharadatunkamsi, 1991) occur on the gested by Kitchener and Maharadatunkamsi’s taxonomic island. In a morphometric evaluation of Indonesian Cyn- revision? Support for this scenario would include two opterus, Kitchener and Maharadatunkamsi (1991) pro- Sulawesi lineages, with sister relationships to either posed that C. brachyotis is replaced on Sulawesi by Philippine, or Sunda shelf C. brachyotis clades. (2) Is there C. luzoniensis and C. minutus, both considered synonyms evidence that two evolutionarily distinct lineages or species of C. brachyotis (Corbet and Hill, 1992; Simmons, 2005), occur currently on the island? Geographic overlap between taxa respectively described from the Philippines and Nias reciprocally monophyletic lineages would provide strong Island, West Sumatra (see Fig. 1). This assessment has support for this scenario. (3) Do T. nigrescens and not been universally accepted (Corbet and Hill, 1992; C. brachyotis exhibit comparable or different levels of Koopman, 1993, 1994, but see Simmons, 2005). While genetic structure within Sulawesi? Based on the endemicity phylogenetic analysis of Cynopterus recovered a single and higher habitat-specificity of T. nigrescens, we expected mitochondrial lineage from Sulawesi (Campbell et al., to find stronger population structure in this species relative 2004) sampling was insufficient to evaluate the Kitchener to the geographically widespread habitat generalist, and Maharadatunkamsi hypothesis. C. brachyotis. 476 P. Campbell et al. / Molecular Phylogenetics and Evolution 44 (2007) 474–482

2. Materials and methods 2.2. Analyses—mitochondrial DNA

2.1. Sampling, PCR, sequencing and fragment analysis Intraspecific phylogenies were reconstructed using maximum parsimony (MP) and Bayesian criteria, implemented Cynopterus brachyotis was sampled from three main- in PAUP* (Version 4.0b10; Swofford, 2002) and MRBAYES land sites on Sulawesi and two offshore islands; Thoopte- (Version 3.1.2; Huelsenbeck and Ronquist, 2001). MP rus nigrescens was sampled from four mainland sites and analyses were run using a full heuristic search with tree- one island (Fig. 1 and Table 1). Total genomic DNA bisection-reconnection (TBR) branch-swapping and ran- was isolated from tissue samples (wing or liver) using Qia- dom stepwise addition (100 replicates). A ca. Seventy-seven gen (Valencia, CA) DNeasy extraction kits. Approxi- base pairs deletion common to all C. brachyotis Sunda mately 640 base pairs (bp) of the 50 end of the haplotypes (see below) was coded as a single synapomor- mitochondrial (mtDNA) control region were sequenced phic character. All other gaps were treated as a fifth char- for C. brachyotis (n = 21; see Table 1 for Genbank Acces- acter state; all characters were weighted equally. Nodal sion Nos.). Primers, PCR conditions and sequencing reac- support was evaluated with 1000 bootstrap replicates, each tion protocols are reported in Campbell et al. (2004); eight with five replicates of random taxon addition, a full heuris- haplotypes from this earlier study were included in the tic search and TBR branch-swapping. present analysis. The C. brachyotis tree was rooted with Megaerops ecaud- For T. nigrescens (n = 37), complete control region was atus, a putative sister genus to Cynopterus (Jones et al., 2002; initially amplified using universal primers (Kocher et al., Giannini and Simmons, 2003). Previously published haplo- 1989). The sequences obtained were used to design internal types from the Philippine-restricted C. brachyotis lineage, primers, ThoopIntF: 50-CCTGAAGTAAGAACCAGAT the Malaysian C. brachyotis Forest lineage, and the wide- G-30 and ThoopIntR: 50-ACGGCATCTGGTTCTTTC-30, spread C. brachyotis Sunda lineage (sensu Campbell et al., which amplified approximately 657 bp at the 50 end of 2004) were included to evaluate the geographic origin of the control region. Amplifications were carried out in a vol- Sulawesi C. brachyotis (see Fig. 2 caption for Genbank ume of 50 ll containing 25 ll of FailSafe PCR 2· PreMix E Accession Nos.). For T. nigrescens, two putative sister taxa (Epicentre, Madison, WI), 0.15 mM of each primer and (Giannini and Simmons, 2003, 2005) were used as out- 0.5 ll of FailSafe PCR enzyme mix. The thermal profile groups: Aethalops alecto and Chironax melanocephalus. (94 C for 30 s; 54 C for 30 s; 72 C for 1 min 30 s) was A likelihood ratio test was implemented in Modeltest repeated for 35 cycles with an initial denaturation step at (Version 3.06; Posada and Crandall, 1998) to find the 94 C for 2 min and a 2 min final extention at 72 C. best-fitting substitution model in each data set. Bayesian PCR products were purified using the QIAquick PCR Puri- analyses were run under the selected GTR + C + I model fication Kit (Qiagen) and sequencing reactions were carried of sequence evolution with default priors and model out using Big Dye terminator (Version 3.1, Applied Biosys- parameters estimated during the course of the run. Four tems). Sequencing reaction products were run on an ABI MCMC chains were run for 2,000,000 generations. Aver- Prism 3730 automated capillary sequencer (Applied Biosys- age log-likelihood values at stationarity were compared tems). Sequences were aligned and edited in SEQUEN- for convergence among chains. Likelihood scores for the CHER (Version 4.0, GeneCodes Corp.). C. brachyotis and T. nigrescens data sets reached stationa- Six microsatellite loci, originally isolated in C. sphinx rity after approximately 200,000 generations; the consensus (CSP-1, -3, -4, -5, -6, and -9; Storz, 2000), amplified in tree saved from the last 1,800,000 generations (post-burn- C. brachyotis (n = 53). CSP-1, -3, -5, -6, and -9 were in) was used to estimate clade position probabilities. retained for analysis as CSP-4 was monomorphic. PCR Network representations of intraspecific genealogical conditions were the same as those described in Campbell relationships can provide an informative alternative to tra- et al. (2006). Amplified products were run on an ABI ditional phylogenetic methods, because the assumption of a Prism 3100 automated capillary sequencer; allele size strictly bifurcating gene tree is relaxed, allowing for persis- was quantified and edited in GeneMapper (Version 3.7). tence of ancestral haplotypes and reticulations resulting A subset of the same loci (CSP-4–6), along with CSP-7, from recombination or homoplasy (Posada and Crandall, were polymorphic in T. nigrescens (n = 38). PCR reactions 2001). This approach is particularly relevant for a locus for T. nigrescens were carried out in a volume of 10 ll such as the control region with a high potential for homo- with 0.5 mM of each primer, 2.5 mM MgCl2, 2.0 mM plasy due to the rapid accumulation of mutations. We con- dNTP’s and 1 U of Taq DNA polymerase (Promega). structed a network for Sulawesi C. brachyotis using the The following thermal profile was repeated for 30 cycles statistical parsimony method (Templeton et al., 1992), with an initial denaturation step at 94 C (2 min) and a implemented in TCS (Version 1.21; Clement et al., 2000) final extension at 72 C (2 min): 94 C (30 s), 55 C with the 10 step criterion and 95% confidence intervals. (30 s), and 72 C (1 min 30 s). PCR products were run Divergence within T. nigrescens was too high to construct on an ABI 377 automated sequencer and fragment data a network for this species with adequate confidence. were extracted and analyzed with GeneScan software Genetic distances among populations and lineages were (Version 3.1.2, Applied Biosystems). calculated in MEGA (Version 3.1, Kumar et al., 2004) under P. Campbell et al. / Molecular Phylogenetics and Evolution 44 (2007) 474–482 477

Table 1 Localities, GenBank Accession Nos. and voucher catalogue numbers for C. brachyotis and T. nigrescens from Sulawesi # Locality, Lat. S Long. E GenBank Accession Nos. Museum cat. #/ Collector ID elevation, n 1 Bangai Dist., 04402500 12300100 C.b., DQ777814–19 T.n., DQ830488 C.b., NK80083, NK80135, NK80137, Siuna, 130 m NK80151, NK80165, NK80173, NK80177; C.b., n =7 T.n., MSB93208 T.n., n =1 2 Poso Dist., 0570 121270 C.b., DQ77820–22 NK80044, NK80048, NK80065, NK80073 Marowo, 10 m C.b., n =3 3 Palu Dist., Lore 011904700 12020400 C.b., DQ77823–24, AY629008,a AY629096,a C.b., PSUT24, PSUT25, PSUT43, PSUT56, Lindu National AY629097,a AY629101,a AY629102,a AY629103a; PSUT195, PSUT196, PDX45, PDX49; T.n., Park, 1050 m T.n., DQ829420, DQ829422, DQ829424, DQ822075, 48LIPI, 51LIPI, 52LIPI, 53LIPI, 85LIPI, C.b., n =8T.n., DQ822079, DQ829710, DQ822067, DQ822099, 123LIPI, 127PSU, 129PSU, 131PSU, n =15 DQ822473, DQ822525, DQ830060, DQ822529, 132LIPI, 133PSU, 134LIPI, 148PSU, DQ822531, DQ822803 160PSU, 161LIPI 4 Sula Arch., 01450000 125500000 T.n., DQ822911 BZM 15232-33 Mangole Isl., <100 m T.n., n =1 5 Central Sulawesi 02540800 1194105000 T.n., DQ822897, DQ822903, DQ822805, DQ830388, NK103513–14, NK103541, NK104009, Prov., Tana DQ830394, DQ822809, DQ830396, DQ830386, NK104016, NK104028–29, NK104041–42, Toraja, 2,150 m DQ830390, DQ830472, DQ830476, DQ822905, NK104049, NK104051–52, NK104054, T.n., n =19 DQ830480, DQ822907, DQ822901, DQ830482, NK104059, NK104061–62, NK104070, DQ830484, DQ830486, DQ 822909 NK104074–75 6 Bau Bau, Buton 051001200 1225001400 C.b., DQ77826–30, AY629094a None Isl., <100 m C.b., n =6 7 S.E. Sulawesi 5250 12200 C.b., DQ77831–32, AY629095a None Prov., Kabaena Isl., <100 m C.b., n =5 8 S. Sulawesi 052302400 1195501200 T.n., DQ822913 NK80014 Prov., G. Lompobatang 1700 m T.n., n =1 C.b., Cynopterus brachyotis; T.n., Thoopterus nigrescens; n, sample sizes for control region sequences; NK, Museum of Southwestern Biology, Albu- querque, NM; PSUT and PDX, Portland State University Museum of Vertebrate Biology, OR; LIPI and PSU specimens deposited at the National Museum of Indonesia, Cibinong; BZM, Bogor Zoological Museum, Indonesia. Locality numbers correspond to locations on maps in Fig. 1. a Campbell et al., 2004. the Tamura–Nei model of nucleotide substitution (Tamura Diversity was estimated by h, which is based on the number and Nei, 1993) using the gamma distributions estimated in of nucleotide polymorphisms (Watterson, 1975), and p, the MrBayes (C. brachyotis, a = 0.44; T. nigrescens, a = 0.30). average pairwise divergence per site (Tajima, 1983). Pairwise estimates of population differentiation based on uncorrected genetic distances among haplotypes (UST) were 2.3. Analyses—microsatellites calculated in ARLEQUIN (Version 3.01; Excoffier et al., 2005) for populations with sample sizes P3. Based on the Tests for linkage disequilibrium and departures from observed lack of mitochondrial or nuclear differentiation Hardy–Weinberg equilibrium were implemented in FSTAT among Cynopterus from Buton and Kabaena Islands, sam- (Version 2.9.3.2; Goudet, 1995). FST was calculated in ples from these sites were treated as a single population. ARLEQUIN and a test for isolation by distance in C. brachyotis Correlation between geographic and genetic distance (iso- was carried out in GENEPOP as for the mtDNA data set, lation by distance) was tested in C. brachyotis using the with linearized FST regressed on geographic distance. We ISOLDE program in GENEPOP (Version 3.4, Raymond and used the genotypic clustering program, STRUCTURAMA Rousset, 1995). Linearized UST was regressed on straight- (Huelsenbeck et al., in press) to further evaluate population line distances between populations (Rousset, 1997). Insuffi- differentiation. STRUCTURAMA uses a Bayesian clustering cient sample sizes precluded this analysis for T. nigrescens. approach to assign individuals to K populations, where the 478 P. Campbell et al. / Molecular Phylogenetics and Evolution 44 (2007) 474–482

Fig. 2. Phylograms representing the consensus trees found using 2,000,000 generations of MCMC sampling in MRBAYES (Huelsenbeck and Ronquist, 2001) for (a) Cynopterus brachyotis and (b) Thoopterus nigrescens, under a GTR + C + I model of nucleotide substitution with parameters estimated during the course of the run. Note that branch lengths for the two phylograms are not on the same scale. Numbers above branches supporting main clades are posterior probabilities; numbers below branches are parsimony bootstrap values based on 100 bootstrap replicates, each with five replicates of random taxon addition, a full heuristic search and TBR branch-swapping. Tip labels correspond to sampling localities, LL, Lore Lindu; TT, Tana Toraja; PM, peninsular Malaysia; Sara, Sarawak; W. Kali, West Kalimantan; numbers in parentheses denote number of haplotypes. Thoopterus nigrescens tree rooted with Aethalops alecto and Chironax melanocephalus (Genbank Accession Nos. AY629149 and AY629150), C. brachyotis tree rooted with Megaerops ecaudatus (AY629151). Genbank Accession Nos. for Philippine, Sunda and Forest C. brachyotis lineages are: AY629024, AY6290047, AY629049, AY629051, AY629066, AY629090, AY629093, AY629099, AY629100, AY629104, AY629105 (Campbell et al., 2004), and AY974394, AY974429, AY974450 (Campbell et al., 2006). posterior probabilities of observing the data given alterna- origins for Cynopterus on Sulawesi, or for strongly differen- tive values of K can be estimated with or without prior tiated lineages within the island. Genetic distance among assumptions of population structure. Individuals are populations was low (mean 2.7% ± SD 0.4%) and both assigned to populations such that the squared distance Bayesian and MP analyses recovered a single, well-sup- among sampled partitions is minimized, where partition dis- ported Sulawesi clade (posterior probability = 0.85, boot- tance is the minimum number of individuals that must be strap = 96). The Bayesian consensus tree is shown in deleted from the assignment vector to make the two parti- Fig. 2a. Differences between Bayesian and MP topologies tions the same (Huelsenbeck et al., in press). Runs of 106 iter- were limited to the positions of haplotypes within the four ations were performed for C. brachyotis and T. nigrescens. C. brachyotis clades included in the analysis (e.g., Sulawesi, Analyses for both species were run using a random variable Philippines, Sunda and Forest). Within the Sulawesi clade, with a Dirichlet process prior for K and prior means of the the only notable difference between the two analyses was number of populations evaluated at 1, 2, 5, and 10. the basal placement of the Lore Lindu subclade under MP, but not Bayesian criteria. Monophyly of both the 3. Results Southeast (Buton/Kabaena) and Lore Lindu subclades was well-supported by Bayesian posterior probability 3.1. Cynopterus brachyotis (0.99 and 0.91) but received low parsimony bootstrap support (57 and 56). A sister relationship between Sulawesi, Mean percent nucleotide diversity in C. brachyotis was and Sunda shelf plus Philippine C. brachyotis was recovered 1.6 (Supplementary Table 1) with 113 parsimony-informa- in both analyses but was poorly supported. The haplotype tive sites. We found no evidence for multiple geographic network constructed under statistical parsimony recovered P. Campbell et al. / Molecular Phylogenetics and Evolution 44 (2007) 474–482 479 the similar relationships within Sulawesi, with Buton/Kaba- terior probability = 0.62). The placement of single haplo- ena haplotypes grouped in one basal cluster, Lore Lindu types from other Sulawesi localities was poorly supported haplotypes in a second cluster and Siuna plus two Poso in both analyses: Gunung Lompobattang and Mangole haplotypes in a third (Supplementary Fig. 1). The Poso hap- Island plus Siuna were basal to the Lore Lindu clade in lotype placed basal to the Lore Lindu clade on the Bayesian the Bayesian analysis and to the Tana Toraja clade under tree (Fig. 2a) could not be placed in the network with 95% parsimony. Genetic distance between Mangole and Siuna confidence. The single reticulation in the network occurred haplotypes was comparatively low (3.7%) with a well-sup- within the Lore Lindu population, suggesting that homo- ported sister relationship in both analyses. plasy is not a significant factor in this data set. We identified 2–9 alleles per microsatellite locus; two We identified 2–9 alleles per microsatellite locus with a alleles were private to Lore Lindu and six were private to small number of private alleles found at each sampling site: Tana Toraja. No linkage disequilibrium was detected two each at Buton/Kabaena and Lore Lindu, one each at between loci but CSP-7 exhibited a significant heterozygote Poso and Siuna. None of the loci exhibited significant link- deficit in the Tana Toraja population (P < 0.001). Both age disequilibrium or departures from Hardy–Weinberg mitochondrial and nuclear differentiation between the Lore equilibrium. Lindu (n = 15) and Tana Toraja (n = 20) populations was There was no correlation between genetic and geo- highly significant, with a UST value of 0.73 for the control 2 graphic distance for either mtDNA (P = 0.3, R = 0.305) region and FST of 0.48 for microsatellites (both, P < 0.0001). 2 or nuclear microsatellites (P = 0.4, R = 0.308). UST values The clustering analysis of microsatellite genotypes for mtDNA ranged from 0.061 to 0.578; all population obtained maximum posterior probability scores for pairs except Poso–Siuna were significantly differentiated K = 2, with all genotypes from the Lore Lindu population (Table 2). Pairwise FST values for microsatellites ranged assigned to cluster 1 and all Tana Toraja genotypes from 0.077 to 0.109, with the Siuna population significantly assigned to cluster 2. Mangole Island and Siuna genotypes differentiated from all other populations (Table 2). How- were assigned to cluster 1; the G. Lompobattang genotype ever, genotypic clustering analysis consistently grouped was assigned to cluster 2. all Sulawesi genotypes in a single cluster with maximum posterior probability estimates obtained for K =1. 4. Discussion

4.1. Taxonomy and phylogeography of Sulawesi C. 3.2. Thoopterus nigrescens brachyotis

Mean percent nucleotide diversity in T. nigrescens was Concordance between mitochondrial and nuclear mark- 1.75 (Supplementary Table 1) with 142 parsimony-infor- ers strongly suggests that the C. brachyotis complex is rep- mative sites. In contrast to C. brachyotis, T. nigrescens resented in Sulawesi by a single lineage. Although we were exhibited considerable genetic structure: mean genetic dis- unable to resolve the biogeographic origin of the Sulawesi tance among sample sites was 7.9%(±SD 1.9%), genetic mitochondrial lineage, recovery of a single well-supported distance between the Tana Toraja and Lore Lindu popula- Sulawesi clade suggests that multiple geographic origins tions was 8.7%. Under both Bayesian and MP criteria, for Sulawesi C. brachyotis are unlikely. haplotypes from Tana Toraja and Lore Lindu were split While a Sulawesi–Philippines dispersal route has been into distinct clades. The Bayesian consensus tree is shown inferred for older taxonomic groups such as birds, butter- in Fig. 2b. Posterior probability and bootstrap support flies (Holloway and Jardine, 1968; Holloway, 1987), frogs for the Lore Lindu clade was high (0.92 and 100, respec- (Emerson et al., 2000; Evans et al., 2003b) and flowering tively). The Tana Toraja clade was well-supported in the plants (Balgooy, 1987), Borneo–Philippines and Borneo– MP (bootstrap = 100), but not the Bayesian analysis (pos- Sulawesi dispersals are implicated in the more recent diversification of mammals (Musser and Heaney, 1992; Ruedi Table 2 et al., 1998; Evans et al., 1999; Lucchini et al., 2005). If Pairwise population differentiation values for Cynopterus brachyotis C. brachyotis colonized Sulawesi via Borneo, transient dis- sampled from four localities in Sulawesi persal opportunities existed from the early Pliocene (ca. 5.2 Population (n) Buton/Kabaena Lore Lindu Poso Siuna million years ago [MYA]) until the end of the Pleistocene (11) (8) (3) (7) (ca. 18,000 years ago), when maximal drops in sea level Buton/Kabaena (12) 0.578*** 0.409** 0.456*** associated with glacial cycles likely reduced the shortest ** *** Lore Lindu (12) 0.072 0.472 0.507 distance across the Makassar Strait to <50 km (Voris, Poso (6) 0.060 0.037 0.061 2000). Support for this hypothesis awaits more rigorous Siuna (23) 0.109*** 0.086** 0.063* sampling of Indonesian C. brachyotis, particularly in East- Mitochondrial UST values and sample sizes (n) are above the diagonal, ern Borneo, and South and North Sulawesi. microsatellite FST values are below with sample sizes in the first column. * P < 0.05. Capture records suggest that the Sulawesi lineage is ** P < 0.01. ecologically similar to the C. brachyotis Sunda and Philip- *** P < 0.001. pine lineages, both of which are most common in disturbed 480 P. Campbell et al. / Molecular Phylogenetics and Evolution 44 (2007) 474–482 habitats (Bergmans and Rozendaal, 1988; Heideman and from the phylogeographic structure of Sulawesi macaques Heaney, 1989; Heaney et al., 1989; Campbell et al., or the Sulawesi toad, sampled from Central Sulawesi sites 2006). However, given that the Sulawesi lineage is on either side of the Palu-Koro fault (Evans et al., restricted to a faunistically unique and geographically iso- 2003a). Likewise, although the single samples from the lated region, it clearly deserves recognition as an evolution- three other sites preclude strong phylogeographic infer- arily significant unit (e.g., Moritz, 1994). Determining ence, control region divergence of P7% among all four whether the Sulawesi Cynopterus lineage warrants species mainland Sulawesi sites, relative to 3.7% between sites sep- status awaits data from more slowly evolving mitochon- arated by a long-standing oceanic barrier (Mangole Island drial and nuclear markers and re-evaluation of relevant and Siuna), suggests a pattern of low gene flow among type specimens. Should taxonomic revision be warranted, Sulawesi T. nigrescens populations, for which vicariance we suggest that Cynopterus minor Revilliod, 1911 (type is one of several potential explanations. Strong population locality: Lambuja, SE Sulawesi) is the available name for structure can also result from aspects of species ecology or the Sulawesian Cynopterus. The findings of this study sup- social behavior (Worthington Wilmer et al., 1994; Irwin, port the view that C. luzoniensis (sensu Kitchener and 2002; Lacey and Wieczorek, 2004). Concordance between Maharadatunkamsi, 1991) is not appropriate, since the nuclear and mitochondrial markers rules out structure type locality associated with the name is in the Philippines, due to sex-biased dispersal or strong male reproductive and Philippine and Sulawesi Cynopterus lineages are recip- skew; however, other factors such as low natal dispersal rocally monophyletic (Campbell et al., 2004). We suggest in both sexes, small home range size or dependence on rare that C. minutus is also inappropriate because it implies lack roost types might also reduce gene flow. of differentiation from Sunda shelf C. brachyotis, a rela- It is also possible that T. nigrescens comprises two or tionship refuted by the monophyly of the Sulawesi mito- more evolutionarily distinct lineages. Numerous other chondrial lineage relative to both C. brachyotis Forest mammal genera have given rise to morphologically and, and Sunda (Fig. 2a). in some cases, ecologically diagnosable species within Sulawesi (macaques, Macaca, Evans et al., 1999; shrews, 4.2. Comparative population structure of Sulawesi Crocidura, Ruedi et al., 1998; rodent genera Taeromys, Cynopterus brachyotis and Thoopterus nigrescens Bunomys, Paruromys, Maxomys,andRattus, Corbet and Hill, 1992; L. Ruedas pers. obs.). While cursory analysis Cynopterus brachyotis is tolerant of human disturbance of general measures of body size found no significant differ- on Sulawesi and belongs to a genus which has successfully ences between Tana Toraja and Lore Lindu T. nigrescens colonized most of Southeast Asia and India, while populations (Putnam, 2004), between-site altitudinal differ- T. nigrescens is a primary forest-associated endemic whose ences of >1000 m (see Table 1) provide ancillary evidence presence on Sulawesi likely predates the arrival of that the two populations may be ecologically differentiated. C. brachyotis. Our results support the prediction that these Determining whether genetic divergence within T. nigres- differences in ecology and evolutionary history should pro- cens on Sulawesi is indicative of low gene flow among pop- duce contrastingly lower genetic structure in C. brachyotis ulations isolated by historic or behavioral barriers, or of relative to T. nigrescens. However, the depth of mitochon- deeper differentiation among demographically and, drial divergence and concordant differentiation at nuclear possibly, ecologically distinct lineages, will require addi- microsatellites in T. nigrescens was unexpected. Notably, tional sampling along both latitudinal and altitudinal gra- the Tana Toraja and Lore Lindu populations in Central dients, and analysis of slower evolving molecular markers Sulawesi are only 180 km apart but exhibit 8.7% control in conjunction with morphometric and ecological data. region divergence, are reciprocally monophyletic (Fig. 2b) In C. brachyotis, the low level of population differentia- and strongly differentiated at nuclear microsatellite loci. tion for microsatellites compared to mtDNA (Table 2) The spatial proximity of the two sites suggests that popula- and lack of genotypic clustering compared to mitochondrial tion differentiation is not an artifact of incomplete geo- monophyly of the geographically isolated Buton/Kabaena graphic sampling. clade (Fig. 2a), are consistent with the smaller effective While no present-day barrier to gene flow exists, the population size and consequently shorter time to dynamic geological history of Sulawesi provides several coalescence for mtDNA (Avise, 2000). Likewise, the lack potential scenarios for vicariance among T. nigrescens pop- of correlation between genetic and geographic distance for ulations. For example, Lore Lindu and Tana Toraja are on either mitochondrial or nuclear markers suggests that opposite sides of the NNW-SSE running Palu-Koro fault C. brachyotis populations have not reached demographic bisecting Central Sulawesi (Bellier et al., 2006; Fig. 1). equilibrium, as is reasonable if colonization is an evolution- The fault marks a suture zone between the Eastern and arily recent event (Slatkin, 1993). It is possible, however, that Western Sulawesi blocks and was a region of high tectonic an effect of isolation by distance would be detected with more activity from the early Miocene through the early Pleisto- thorough geographic sampling (e.g., Bridle et al., 2004). cene (ca. 23–1.5 MYA), including a period of major mon- Taken together, the results of this study contribute to tane uplift in the early Pliocene (ca. 5 MYA; Hall, 2002). knowledge of the diversity of Old World fruit bats in a We note, however, that no such barrier has been inferred unique biogeographic region where prior studies of this P. Campbell et al. / Molecular Phylogenetics and Evolution 44 (2007) 474–482 481 group have been limited to inclusion of single samples in Bridle, J.R., Pedro, P.M., Butlin, R.K., 2004. Habitat fragmentation and deep phylogenies (Colgan and Flannery, 1995; Colgan biodiversity: testing for evolutionary effects of refugia. Evolution 58, and da Costa, 2002; Giannini and Simmons, 2003, 2005). 1394–1396. Campbell, P., Schneider, C.J., Adnan, A.M., Zubaid, A., Kunz, T.H., Our finding of a monophyletic Cynopterus lineage in 2004. Phylogeny and phylogeography of Old World fruit bats in the Sulawesi highlights the importance of comparing multiple Cynopterus brachyotis complex. Molecular Phylogenetics and Evolu- lines of evidence when redefining species distributions or tion 33, 764–781. renaming species. High differentiation within T. nigrescens Campbell, P., Schneider, C.J., Adnan, A.M., Zubaid, A., Kunz, T.H., suggests that further molecular studies of this species, and 2006. Comparative population structure of Cynopterus fruit bats in peninsular Malaysia and Southern Thailand. Molecular Ecology 15, of other endemic bats in Sulawesi, are likely to reveal addi- 29–47. tional unrecognized diversity. Clement, M., Posada, D., Crandall, K., 2000. TCS: a computer program to estimate gene genealogies. Molecular Ecology 9, Acknowledgments 1657–1660. Colgan, D., Flannery, T., 1995. A phylogeny of Indo-West Pacific megachiroptera based on ribosomal DNA. Systematic Biology 44, This manuscript was significantly improved by the com- 209–220. ments of two anonymous reviewers. We thank Ben Evans Colgan, D., da Costa, P., 2002. Megachiropteran evolution studied with and John Huelsenbeck for helpful discussions, Rasit Bilgin 12s rDNA and c-mos DNA sequences. Journal of Mammalian for T. nigrescens internal primer design, and Tigga Kings- Evolution 9, 3–32. ton and Steve Rossiter for generous donation of Cynopte- Corbet, G.B., Hill, J.E., 1992. The Mammals of the Indomalayan Region: A Systematic Review. British Museum Publications, Oxford Univer- rus tissue samples from Buton and Kabaena Islands. sity Press, Oxford, UK. Portions of this work were supported by NSF grant IOB Emerson, S.B., Inger, R.F., Iskandar, D.T., 2000. Molecular systematics (DEB) 0075555 to JCM and LAR, and National Geo- and biogeography of the fanged frogs of Southeast Asia. Molecular graphic Foundation Grant No. 6177–98 to LAR; CB was Phylogenetics and Evolution 16, 131–142. supported by an Undergraduate Research Opportunities Evans, B.J., Morales, J.C., Supriatna, J., Melnick, D.J., 1999. The origin of the Sulawesi macaques (Cercopithecidae, Macaca) as inferred from Program award from Boston University. Research and col- mitochondrial DNA phylogeny. Biological Journal of the Linnean lection permits for this work were granted by the Indone- Society 66, 539–560. sian Institute of Sciences (LIPI); export permits were Evans, B.J., Supriatna, J., Andayani, N., Setiadi, M.I., Cannatella, D.C., granted by the Indonesian Ministry of Forestry (CITES of- Melnick, D.J., 2003a. Monkeys and toads define areas of endemism on fice) under authority of the National Museum of Indone- the island of Sulawesi. Evolution 57, 1436–1443. Evans, B.J., Brown, R.M., McGuire, J.A., Supriatna, J., Andayani, N., sia, Cibinong (Bogor Museum). We thank our field Diesmos, A., Iskandar, D.T., Melnick, D.J., Cannatella, D.C., 2003b. companions R.M. Brown (University of Kansas) and Phylogenetics of fanged frogs (Anura; Ranidae; Limnonectes): testing R.W. Dickerman (University of New Mexico), and our biogeographical hypotheses at the Asian–Australian faunal zone Indonesian colleagues Noviar Andayani, Jatna Supriatna, interface. Systematic Biology 52, 794–819. Nurul Fadhila, and Nenny Babo for their invaluable Evans, B.J., Cannatella, D.C., Melnick, D.J., 2004. Understanding the origins of areas of endemism in phylogeographic analyses: a reply to assistance. Bridle et al. Evolution 57, 1397–1400. Excoffier, L., Laval, G., Schneider, S., 2005. ARLEQUIN version 3.0: an Appendix A. Supplementary data integrated software package for population genetics data analysis. Evolutionary Bioinformatics Online 1, 47–50. Supplementary data associated with this article can be Giannini, N.P., Simmons, N.B., 2003. A phylogeny of megachiropteran bats (Mammalia: Chiroptera: Pteropodidae) based on direct optimi- found, in the online version, at doi:10.1016/ zation analysis of one nuclear and four mitochondrial genes. Cladistics j.ympev.2007.02.013. 19, 496–511. Giannini, N.P., Simmons, N.B., 2005. Conflict and congruence in a combined DNA-morphology analysis of megachiropteran bat rela- References tionships (Mammalia: Chiroptera: Pteropodidae). Cladistics 21, 411–437. Avise, J.C., 2000. Phylogeography: The History and Formation of Species. Goudet, J., 1995. FSTAT version 1.2: a computer program to calculate Harvard University Press, Cambridge, Massachusetts. Fstatistics. Journal of Heredity 86, 485–486. Balgooy, M.M.J., 1987. A plant biogeographical analysis of Sulawesi. In: Hall, R., 2002. Cenozoic geological and plate tectonic evolution of SE Whitmore, T.C. (Ed.), Biogeographical Evolution of the Malay Asia and the SW pacific. Journal of Asian Earth Sciences 20, Archipelago. Clarendon Press, Oxford, UK, pp. 94–102. 341–353. Bellier, O., Se´brier, M., Seward, D., Beaudouin, T., Villeneuve, M., Heaney, L.R., Heideman, P.D., Rickart, E.A., Utzurrum, R.B., Klompen, Putranto, E., 2006. Fission track and fault kinematics analyses for new J.S.H., 1989. Elevational zonation of mammals in the central Philip- insight into the Late Cenozoic tectonic regime changes in West-Central pines. Journal of Tropical Ecology 5, 259–280. Sulawesi (Indonesia). Tectonophysics 413, 201–220. Heideman, P.D., Heaney, L.R., 1989. Population biology and Bergmans, W., Rozendaal, F.G., 1988. Notes on collections of fruit bats estimates of abundance of fruit bats (Pteropodidae) In Philippine from Sulawesi and some off-lying islands (Mammalia, Megachirop- submontane rainforest. Journal of Zoology (London) 218, tera). Zoologische Verhandelingen 248, 1–74. 565–586. Bridle, J.R., Garn, A.K., Monk, K.A., Butlin, R.K., 2001. Speciation in Holloway, J.D., 1987. Lepidoptera patterns involving Sulawesi: what do Chitaura grasshoppers (Acrididae: Oxyinae) on the island of Sulawesi: they indicate of past geography? In: Whitmore, T.C. (Ed.), Biogeo- color patterns, morphology and contact zones. Biological Journal of graphical Evolution of the Malay Archipelago. Clarendon Press, the Linnean Society 72, 373–390. Oxford, UK, pp. 103–118. 482 P. Campbell et al. / Molecular Phylogenetics and Evolution 44 (2007) 474–482

Holloway, J.D., Jardine, N., 1968. Two approaches to zoogeography: a Raymond, M., Rousset, F., 1995. GENEPOP Version 1.2: population study based on the distributions of butterflies, birds and bats in the genetics software for exact tests and ecumenicism. Journal of Heredity Indo-Australian area. Proceedings of the Linnean Society of London 86, 248–249. 179, 153–188. Revilliod, P., 1911. Uber einige Saügeteire von Celebes. Zoologischer Huelsenbeck, J.P., Ronquist, F., 2001. MRBAYES: Bayesian inference of Anzeiger 37, 513–517. phylogenetic trees. Bioinformatics 17, 754–755. Rousset, F., 1997. Genetic differentiation and estimation of gene flow from Huelsenbeck, J.P., Andolfatto, P., Huelsenbeck, E.T., in press. Structu- F-statistics under isolation by distance. Genetics 145, 1219–1228. rama: bayesian inference of population structure. Bioinformatics. Ruedi, M., 1995. Taxonomic revision of Crocidura shrews from the Sunda Irwin, D., 2002. Phylogeographic breaks without geographic barriers to shelf and Sulawesi with description of two new species (Mammalia: gene flow. Evolution 56, 2383–2394. Soricidae). Zoological Journal of the Linnean Society 115, 211–265. Jones, K.E., Purvis, A., MacLarnon, A., Bininda-Emonds, O.R.P., Ruedi, M., Auberson, M., Savolainen, V., 1998. Biogeography of Simmons, N.B., 2002. A phylogenetic supertree of the bats (Mamma- Sulawesian shrews: testing for their origin with a parametric bootstrap lia: Chiroptera). Biological Reviews 77, 223–259. on molecular data. Molecular Phylogenetics and Evolution 9, 567–571. Kitchener, D.J., Maharadatunkamsi, 1991. Description of a new species of Simmons, N.B., 2005. Order Chiroptera. In: Wilson, D.E., Reeder, D.M. Cynopterus (Chiroptera: Pteropodidae) from Nusa Tenggara Indone- (Eds.), Mammal Species of the World: A Taxonomic and Geographic sia. Records of the Western Australian Museum 51, 307–363. Reference, third ed. Johns Hopkins University Press, Baltimore, Kocher, T.D., Thomas, W.K., Meyer, A., Edwards, S.V., Paä¨bo, S., Maryland, pp. 312–529. Villablanca, F.X., Wilson, A.C., 1989. Dynamics of mitochondrial Slatkin, M., 1993. Isolation by distance in equilibrium and non-equilib- DNA evolution in animals—amplification and sequencing with con- rium populations. Evolution 47, 264–279. served primers. Proceedings of the National Academy of Sciences of Storz, J.F., 2000. Variation at tri- and tetranucleotide repeat microsatellite the United States of America 86, 6196–6200. loci in the fruit bat genus Cynopterus (Chiroptera: Pteropodidae). Koopman, K.F., 1993. Order Chiroptera. In: Wilson, D.E., Reeder, D.M. Molecular Ecology 9, 2198–2200. (Eds.), In Mammal Species of the World: A Taxonomic and Swofford, D.L., 2002. PAUP*, Phylogenetic analysis using parsimony (*and Geographic Reference, second ed. Smithsonian Institution Press, other methods). Version 4.0b10. Sinauer Associates, Sunderland, Washington, DC, pp. 137–241. Massachusetts. Koopman, K.F., 1994. Chiroptera: Systematics. Handbook of Zoology: A Tajima, F., 1983. Evolutionary relationship of DNA sequences in finite Natural History of the Phyla of the Animal Kingdom. VIII Mamma- populations. Genetics 105, 437–460. lia. Walter de Gruyter, New York. Tamura, K., Nei, M., 1993. Estimation of the number of nucleotide Kumar, S., Tamura, K., Nei, M., 2004. MEGA3: Integrated software for substitutions in the control region of mitochondrial DNA in humans Molecular Evolutionary Genetics Analysis and sequence alignment. and chimpanzees. Molecular Biology and Evolution 10, 512–526. Briefings in Bioinformatics 5, 150–163. Templeton, A.R., Crandall, K.A., Sing, C.F., 1992. A cladistic analysis of Lacey, E.A., Wieczorek, J.R., 2004. Kinship in colonial tuco-tucos: phenotypic associations with haplotypes inferred from restriction evidence from group composition and population structure. Behav- endonuclease mapping and DNA sequence data. III. Cladogram ioral Ecology 15, 988–996. estimation. Genetics 132, 619–633. Lucchini, V., Meijaard, E., Diong, C.H., Groves, C.P., Randi, E., 2005. Voris, H.K., 2000. Maps of Pleistocene sea levels in Southeast Asia: New phylogenetic perspectives among species of South-east Asian wild shorelines, river systems and time durations. Journal of Biogeography pig (Sus sp.) based on mtDNA sequences and morphometric data. 27, 1153–1167. Journal of Zoology (London) 266, 25–35. Wallace, A.R., 1876. The Geographical Distribution of Animals. Mac- Moritz, C., 1994. Defining evolutionarily significant units for conserva- millan, London. tion. Trends in Ecology and Evolution 9, 373–375. Walton, C., Butlin, R.K., Monk, K.A., 1997. A phylogeny for grasshop- Musser, G.G., Heaney, L.R., 1992. Philippine rodents—definitions of pers in the genus Chitaura (Orthoptera: Acrididae) from Sulawesi, Tarsomys and Limnomys plus a preliminary assessment of phylogenetic Indonesia, based on mitochondrial DNA sequence data. Biological patterns among native Philippine murines (Murinae, Muridae). Bul- Journal of the Linnean Society 62, 365–382. letin of the American Museum of Natural History 211, 2–138. Watterson, G.A., 1975. On the number of segregating sites in genetical Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B., models without recombination. Theoretical Population Biology. 7, Kent, J., 2000. Biodiversity hotspots for conservation priorities. 256–276. Nature 403, 853–858. Whitten, T., Whitten, J., Mittermeier, C.G., Suprianta, J., Mittermeier, Posada, D., Crandall, K.A., 1998. MODELTEST: testing the model of R.A., 1997. Indonesia. In: Mittermeier, R.A., Gil, P.R., Mittermeier, DNA substitution. Bioinformatics 14, 817–818. C.G. (Eds.), Megadiversity: Earth’s Biologically Wealthiest Nations. Posada, D., Crandall, K.A., 2001. Intraspecific gene genealogies: trees Cemex, Mexico City, Mexico, pp. 74–107. grafting into networks. Trends in Ecology and Evolution 16, 37–45. Worthington Wilmer, J., Moritz, C., Hall, L., Toop, J., 1994. Extreme Putnam, A., 2004. Population genetics of the swift fruit bat, Thoopterus population structuring in the threatened ghost bat, Macroderma gigas. nigrescens in Sulawesi. MS thesis, Columbia University, New York. Proceedings of the Royal Society of London, Series B 257, 193–198.